The Measurement of Customer Satisfaction

Table of Contents

Figures

Tables

Acronyms

1 Introduction

2 Theoretical Framework
2.1 Formation of Customer Satisfaction
2.1.1 The C/D-Paradigm as a Conceptual Framework
2.1.2 In-Depth Theories of Satisfaction Formation
2.2 Consequences of Customer Satisfaction
2.3 Overview of Measurement Methodology
2.3.1 Objective vs. Subjective Methods
2.3.2 Event- vs. Attribute-Specific Methods
2.3.3 Indirect vs. Direct Measurement
2.3.4 One-Dimensional vs. Multi-Dimensional Methods
2.3.5 Ex Ante/Ex Post vs. Ex Post Measurement

3 Subjective Measurement Methods
3.1 Event-Specific Measures
3.1.1 Critical Incident Technique
3.1.2 Frequency Relevance Analysis for Problems
3.2 Attribute-Specific Measures
3.2.1 SERVQUAL
3.2.2 SERVPERF
3.2.3 SERVIMPERF
3.2.4 Customer Satisfaction Index
3.2.5 National Customer Satisfaction Indices
3.3 Comparison and Critical Evaluation

4 Comparability of Customer Satisfaction Measures
4.1 Sources of Error for International Comparisons
4.2 Single- vs. Multi-Dimensional Measures ‑ An Empirical Study
4.2.1 Data and Study Approach
4.2.2 Results and Discussion

5 Customer Satisfaction across Industries and Countries
5.1 Comparison of National Customer Satisfaction Indices
5.1.1 The American Customer Satisfaction Index (ACSI)
5.1.2 The German Customer Satisfaction Barometer
5.1.3 The Extended Performance Satisfaction Index (EPSI)
5.2 Cross-Industrial Findings
5.3 Cross-National Findings

6 Conclusion and Outlook

Bibliography

Figures

Figure 1: The C/D-Paradigm: Relationships and Subsumption of Theoretical Concepts

Figure 2: Satisfaction and Customer Experience

Figure 3: Overview of Customer Satisfaction Measurement Methods

Figure 4: Examples of Ex Ante/Ex Post and Ex Post Surveys of Customer Satisfaction

Figure 5: Example of a Pareto Diagram used in FRAP

Figure 6: Measurement Criteria of Attribute-Specific Methods

Figure 7: Importance-Performance Analysis

Figure 8: Psychological Effect of Different Scales

Figure 9: Example of the Calculation of the CSI as Arithmetic Mean

Figure 10: Illustration of the Causal Model of the ACSI

Figure 11: Course of Action for a Holistic Measurement of Customer Satisfaction and Loyalty

Figure 12: The Problem of Non-Equivalent Anchor Points

Figure 13: Absolute Values of Customer Satisfaction Indices for ACSI and BrandIndex

Figure 14: Rescaled CSI-Levels

Figure 15: Illustration of the Causal Model of the EPSI

Figure 16: EPSI 2007 Customer Satisfaction for the Banking Sector, by Country

Tables

Table 1: Overview of In-Depth Theories and Concepts of Satisfaction
Formation

Table 2: Comparative Assessment of the Customer Satisfaction Measurement Approaches

Table 3: Objectives of National Customer Satisfaction Barometers

Acronyms

illustration not visible in this excerpt

1 Introduction

Only those companies that fully satisfy their customers will be able to maintain a top market position in the long run. Despite the fact that this rule has been known for a very long time, it has drastically gained in importance in the last years. The reasons are multifarious, but three main aspects can be named: the upswing of customer needs, exacerbating competition, and more complex market conditions. Rising customer demands can best be explained by a growing range of products competing against each other.

In times of fierce competition, companies must deliver optimal products for the purpose of maintaining their competitiveness and strengthening their market positions. To ensure maximum customer loyalty, it is essential to know the status quo regarding the so called customer satisfaction and how consumer needs developed over time.

Customer satisfaction is described as a construct of the social sciences in the literature. In most cases, it is seen as the relationship of a customer’s expectations and their fulfillment. Satisfaction is a feeling that arises when a product or service meets or surpasses the customer’s expectation. But in order to get a grasp of this abstract construct, scientific and yet practical instruments have to be developed. These tools need to provide fast and detailed results that allow for analyzing different components of satisfaction.

But, over time, the number and complexity of influencing factors has grown. Nowadays, not only needs for the domestic market have to be considered. Most firms act in international markets where the consumer expectations vary considerably. Reasons for this can be of cultural, financial, and even geographic nature amongst other things. A method for assessing customer satisfaction, thus, also has to meet the requirement of international comparability.

Even now, after 25 years of intensive practical application, there are no signs of diminishing interest in measuring customer satisfaction. One must rather state a significant increase in spending, despite the strained economic situation. In the process, the professionalism in executing measurements and implementing appropriate actions has improved considerably. However, there is still not the one way of measuring customer satisfaction. Different courses of action exist depending on country, industrial sector, institute, and field of application. Consequently, a strong demand for an illustration of the different concepts of customer satisfaction and their corresponding measurement methods persists.

This thesis presents the current state of research in the field of customer satisfaction measurement. In its macro-structure it can be divided into a theoretical and an empirical part. In the first one, the main measurement and calculation methods are described and discussed with focus on their capability to provide valid and reliable results concerning customer satisfaction. Chapter 2 sets the thematic framework by examining different conceptions of the formation of customer satisfaction. On this basis, a coarse presentation and categorization of prevalent measurement approaches is given. Not all of those approaches are up to today’s standard of providing valid and reliable measures. Hence, in Chapter 3, only the promising subset of those approaches, the so-called subjective methods, is further analyzed. These can be distinguished into event- (Chapter 3.1) and attribute-specific (Chapter 3.2) methods. Those methods are critically analyzed and assessed with the help of the existing literature. Building on this, a holistic customer satisfaction measurement system is introduced in Chapter 3.3. Subsequently, Chapter 4 addresses the issue of comparability of different customer satisfaction measures and represents the transition from the theoretical to a praxis-based empirical part. Chapter 4.1 focuses on challenges for obtaining comparable data in international measurements. An empirical study is conducted in the second sub-chapter to determine the suitability of one-dimensional surveys as indicators for the prevailing multi-dimensional national CSI models. Hence, the survey data of an attribute-specific approach, the American Customer Satisfaction Index ACSI, are compared with the data of a one-dimensional, attribute-specific approach, the YouGov BrandIndex. A closer look is taken at the different national CSI models as well as a recently established uniform European approach, the Extended Performance Satisfaction Index EPSI (Chapter 5.1) afterwards. Furthermore, in Chapter 5.2 and 5.3, study results of individual countries and industries are examined for their similarities and differences. The thesis concludes with a summary of the most important findings and provides an outlook on future developments in the field of customer satisfaction research.

2 Theoretical Framework

Customer satisfaction takes a central role in current marketing theory and praxis. As such, it constitutes an important link between customers’ behaviors and the activities of a company. Customer satisfaction, correspondingly, has been the object of research in far more than 15.000 studies; only counting works from the USA^[1]. As a result, a variety of theories was developed to explain this abstract construct. To this date, however, consensus on that matter has yet to be reached^[2].

This chapter examines different conceptions of the formation of customer satisfaction. For this purpose, the C/D-paradigm is presented as a baseline framework in which more specific theories can be integrated. Subsequently, a brief presentation and evaluation of existing methods for measuring customer satisfaction sets the focus for the following chapters.

2.1 Formation of Customer Satisfaction

In order to systemize theories and concepts which are relevant to the development of customer satisfaction, the C/D-paradigm (Confirmation/ Disconfirmation-Paradigm) is used. This concept has reached broad support within the context of customer satisfaction measurement^[3]. Hence, its four main components are defined and discussed below. The C/D-paradigm alone, however, does not sufficiently describe the relationships between them. Therefore, additional concepts are presented to complement the overall picture of the formation of customer satisfaction.

2.1.1 The C/D-Paradigm as a Conceptual Framework

The quintessence of the C/D-paradigm is that satisfaction results from the comparison of the perceived actual performance of a product or service and a pre-consumptive comparison standard of the targeted performance. If those two performances match, confirmation occurs, while mismatches lead to either positive or negative disconfirmation, depending on which factor exceeds the other one^[4]. Since a meta-analysis based on 50 studies was conducted by Szymanski/Henard, an additional direct effect of perceived performance on satisfaction is widely accepted^[5]. That means performance has a double impact on customer satisfaction: in a direct way on the one hand and indirectly through disconfirmation on the other hand. A similar direct effect for expectations, however, could not be verified.^[6]

Furthermore, perceived performance and comparison standard do not seem to be autonomous, but mutually depending on each other. Given a discrepancy between these two parameters, a belated correction of the expectations or the perceived performance, respectively, is in fact supposed to take place. Following this correction, the result is either a decrease or an increase of the disconfirmation depending on its initial extent. That is the object of investigation of the Assimilation Theory and the Contrast Theory presented further below.

illustration not visible in this excerpt

Figure 1: The C/D-Paradigm: Relationships and Subsumption of Theoretical Concepts

(Own illustration, loosely based on Anderson, E.W., Sullivan, M.W. (1993), p. 127, Churchill, G.A., Surprenant, C. (1982), p. 426; Homburg, C. et al. (1999), p. 176.)

Figure 1 illustrates the C/D-paradigm with its four main components

(1) comparison standard,
(2) perceived performance,
(3) disconfirmation,
(4) and satisfaction

as well as theories aiming to explain their relationships.

The comparison standard reflects the level of a customer’s expectation towards a product or service. In the literature, especially the following dimensions are discussed^[7]:

- Expectations
- Experience
- Ideals

While expectations relate to an anticipated level of performance^[8], experience is based on previous encounters with the same or similar products^[9]. The optimal performance level possible is taken as a basis when customers use ideals as the comparison standard. It is also assumed that customers can have more than one comparison standard to reach an opinion about their satisfaction. Tse/Wilson could, for example, prove that three different standards influence customers not only simultaneously but sequentially as well^[10]. In addition, it is suggested that the type and intensity of the perception of comparison standards vary within different situations. Johnson even argues that with growing product experience, expectations and perceived performance become increasingly integrated^[11]. Figure 2 illustrates three different stages of experience and their impact on satisfaction.

illustration not visible in this excerpt

Figure 2: Satisfaction and Customer Experience

(Own illustration, based on Johnson, M.D., Fornell, C. (1991), p. 11.)

Using a product for the first time, consumers cannot rely on previous memories or accumulated experiences. Thus, expectations are not considered to have a significant effect on satisfaction or perceived performance. The more encounters customers have had with a product, the better they are able to estimate the performance of it and, this way, close the information gap between these two constructs.

Performance of a product or service, in contrast to expectations, attracts only minor interest in the literature. A distinction is made between objective and subjective performance^[12]. While objective (that is actual) performance is the same for all customers, subjective performance varies due to several effects on their individual perception. Within the scope of operationalizing perceived performance, subjective performance is preferred in the literature.

Disconfirmation is seen as the central variable among performance, expectations, and the resulting satisfaction^[13]. Concerning this comparison, the focus lies on the ratio between the performance and the comparison standard, with its three outcomes for disconfirmation as described above. In order to measure the difference between the perceived and the targeted performance, researchers have commonly captured the targeted and the perceived performance separately. Afterwards, the difference was calculated by mathematical subtraction^[14]. This approach has received criticism due to its inability to capture individual differences^[15]. Therefore, the subjective discrepancy between perceived and expected performance is increasingly measured directly in the recent literature^[16].

Satisfaction, the last variable in the C/D-paradigm, is in many cases defined as the result of a cognitive comparison^[17]. Whereas this understanding of satisfaction was predominant in scientific publications for a long time, newer works base their satisfaction construct on a broader concept. Alongside the cognitive component, an affective component is included into the definition of customer satisfaction^[18]. The relevance of the emotional state of a consumer could be demonstrated in numerous studies^[19].

2.1.2 In-Depth Theories of Satisfaction Formation

As already stated, the C/D-paradigm serves as a baseline framework that provides an idea of how the relevant variables interact in general. Other theories, however, can be integrated into this concept to explain the formation of customer satisfaction in a more detailed way.

The Assimilation Theory, the Contrast Theory and the Assimilation-Contrast Theory concentrate on the ex post alteration of the comparison standard or the perceived performance, respectively, in case of disconfirmation. In addition, several approaches examine the relationship between disconfirmation and the degree of satisfaction. While the basic model of the C/D-paradigm assumes a certain amount of confirmation or disconfirmation always leading to a certain extent of satisfaction, Attribution Theory indicates that subjective perception influences this relationship. In addition, the different natures of a product’s attributes significantly impact a consumer’s satisfaction, as stated by the Kano Model. Prospect Theory, ultimately, explains why the relation from disconfirmation to satisfaction cannot be seen as linear. Table 1 provides an outline of those theories including their core statements as well as a selection of relevant authors.

illustration not visible in this excerpt

Table 1: Overview of In-Depth Theories and Concepts of Satisfaction Formation

2.2 Consequences of Customer Satisfaction

Like any other investment, money spent on activities in order to increase customer satisfaction should eventually yield an economic advantage. Thus, a large number of studies have been conducted to investigate the influence of customer satisfaction on consumer behavior and, more directly, on economic performance. The bigger part of those studies focuses on the link between customer satisfaction and loyalty. Most definitions break customer loyalty down to three dimensions: repeat purchasing, cross-buying, and word-of-mouth recommendation^[20]. A positive influence of satisfaction on these aspects is generally held to be true. But while there is no doubt that a strong and positive link between customer satisfaction and loyalty exists^[21], the exact relation is still unclear and probably more complex than previously imagined^[22]. Different empirical studies claim the relationship to be progressive^[23], saddle-shaped^[24], declining^[25] or S-shaped^[26], respectively.

Another aspect of consumer behavior is a customer attitude to price. Research in this area, however, is still in its beginning. Whilst most scholars’ conclusions, that satisfaction has a positive influence on a customer’s willingness to pay, only rely on plausible considerations, Anderson’s empirical analysis of data from Sweden achieved mixed results^[27]. A static view shows a negative relationship between customer satisfaction and price tolerance. Using a dynamic view over two years, on the other hand, results in a positive relationship. This suggests a more long-term perspective on the efficacy of actions relating to customer satisfaction.

While the former studies concentrated on the consumer side, increasing efforts are made to directly measure the economic returns of customer satisfaction. Results clearly prove a positive relationship, measured for example by ROI^[28], ROA^[29], and Tobin’s Q^[30]. Additionally, a positive relationship between customer satisfaction and cash flow as well as a negative one between customer satisfaction and the volatility of the cash flow could be proven^[31]. But, more is not always better. Fischer et al. show that there is no simple linear relationship of customer satisfaction and economic performance of a firm^[32]. Rather, spending too much money on customer satisfaction can result in decreasing profits or even losses, when the optimal level is exceeded. This necessarily leads to the question: How can customer satisfaction be measured effectively in order to allocate a firm’s resources in a way that maximizes the outcome?

2.3 Overview of Measurement Methodology

In the literature, the classification of customer satisfaction measurement methods by Andreasen has established itself and was refined over the years^[33]. According to it, measurement methods are primarily classified as objective or subjective. Subjective methods can furthermore be broken down into several dimensions. Figure 3 serves as an orientation for the approaches discussed in detail below.

illustration not visible in this excerpt

Figure 3: Overview of Customer Satisfaction Measurement Methods

(Own illustration, based on Koch, M. et al. (2004), pp. 189.)

2.3.1 Objective vs. Subjective Methods

Objective methods capture customer satisfaction with the help of observable quantities that are not influenced by the investigator’s interpretation. Among them, there are monetary quantities like market share, sales volume, EBIT, etc. And although studies show a positive link between customer satisfaction and these economic figures, there have not been any scientifically backed findings about a functional relationship^[34]. This is due to the problematic that these figures are also influenced by several unpredictable external factors like economic cycles, exchange rates, and competitor activities, to mention a few. The subjective satisfaction of a consumer may be very high for example, although EBIT and sales volume decline.

As an alternative, mystery shoppers and product tests by experts or independent institutions are discussed. But those test purchases cannot reflect the emotions that a real customer has buying a product. Additionally, test purchases are judged more critically most of the time^[35]. In summary, objective methods are not considered to be suitable for a valid and reliable measurement of customer satisfaction because of the influence of other external as well as internal factors. For those reasons, this work will from here on concentrate on subjective methods only.

Subjective methods, contrariwise, rely on a pre-defined construct of customer satisfaction and attempt to measure it via indicators. These can be based on physical or psychological circumstances. Only customer surveys can assure product or service are in line with the market’s needs and make the customers’ assessment of value more transparent^[36]. Depending on the object of investigation, subjective methods are classified as either event- or attribute-specific.

2.3.2 Event- vs. Attribute-Specific Methods

Event-specific methods are based on transactions or incidents a customer experienced with the firm, so called “moments of truth”. Those moments contain all kinds of routine and non-routine customer experiences such as a car workshop visit or bringing a new tooling machine into service for an industry customer^[37]. Most methods focus on important events, though, because they are stored in memory for a longer period of time. And timing is a critical factor with event-specific methods because, on the one hand, customers need a certain amount of time to form an opinion of their recent encounter. On the other hand, the exact reminiscence of the event may fade as time passes.

Concentrating on certain incidents of a customer contact is certainly helpful with regard to a selective improvement within the company, especially for services. For a comprehensive view on customer satisfaction, however, event-specific methods are not well suited due to their relatively small period under review. They can comprise only one or a few single events and therefore cannot cover the whole extent of a customer relationship^[38]. Qualitative research showed, though, that consumers distinguish between transaction-specific and overall concepts of satisfaction. That makes both approaches reasonable measures depending on the context. Common event-specific methods are the Critical Incident Technique and the Frequency Relevance Analysis for Problems which will be presented in Chapter 3.

Attribute-specific methods deal with a broad spectrum of product-, service- or interaction-attributes about which the customer forms an opinion over time. This process can go on for years, fortifying or changing this opinion. For this reason, those methods are also known as cumulative approaches. Attribute-specific methods are suited for cost-effective, timely and standardized measurements because the desired attributes can be queried within the scope of an extensive customer survey and aggregated to a measure of customer satisfaction^[39]. The different concrete approaches can be subdivided in terms of their directness of measurement^[40].

2.3.3 Indirect vs. Direct Measurement

Indirect measurements infer customer satisfaction from its consequences. Normally, perceived performance deficits are derived from a complaint analysis. The critical assumption is that high satisfaction leads to fewer complaints. In the majority of cases, the complaints are accumulated, individually handled, but centrally evaluated in the end. This allows drawing conclusions about customer satisfaction, as a rise in complaints is interpreted as an indicator for a problem.

A major disadvantage of indirect measurement is the fact that it requires active complaining behavior of the customers. Lacking chances of success, the trouble associated with the complaint, and the expenditure of time cause customers to refrain from a complaint^[41]. Thus, an absence of complaints cannot necessarily be attributed to satisfaction. There is also the chance of silent customer defection. In the case of a bank, a client might still have his account open, just leaving a small amount of money on it, while already having an account at another bank. The same is true for the ordering of spare parts for a machine that a customer is not satisfied with. Nevertheless, a professional complaint management is a good way to keep loyal and recover lost ones, and several techniques can improve the customers’ complaint behavior^[42]. However, the main points of criticism, that this method of measuring customer satisfaction is often unsystematic and not always representative, remain.

Using a direct measurement, customers are surveyed firsthand about their satisfaction or the fulfillment of their expectations with the help of one- or multi-dimensional scales^[43]. Designing a survey includes several areas of decision, most notably:

- Target group
- Sample design
- Sort of the survey
- Way of questioning
- Choice of Measurement scale
- Subject matter of the survey
- Anonymity of the survey
- Timing of the survey

Direct measurements can be sub-classified by their dimensionality, being either one- dimensional or multi-dimensional.

2.3.4 One-Dimensional vs. Multi-Dimensional Methods

If customer satisfaction is measured by only one area of satisfaction, the method used is one-dimensional. Often, customers are only asked for the overall satisfaction with a firm’s performance. This global statement, however, makes it impossible to determine the reasons for this judgment. Additionally, it remains unclear, if the answer is based on a rational thought process and follows a functional relationship or if it is a rather intuitive, across-the-board response. The advantages of those methods are rooted in their simple handling and their low complexity. Yet, they do not analyze the satisfaction with sufficient differentiation^[44]. Due to the fact that those methods capture the complex construct of satisfaction with only one indicator, there is an increasing consensus within the scientific community that they are insufficient in terms of reliability and validity^[45]. Moreover, the diversity of products and services offered in praxis requires the differentiated analysis of single performance factors. While a one-dimensional method cannot offer this, the measured overall satisfaction can be a valuable benchmark to the results of a multi-dimensional measure. In the context of a multiple regression, it serves as an important basis in order to quantify the contribution of the individual criteria to the total satisfaction. In Chapter 4, an empirical study is conducted to test the suitability of one-dimensional measures as indicators for multi-dimensional ones with a concrete example.

Multi-dimensional methods collect a composition of additive judgments of performance criteria and calculate an overall satisfaction rating^[46]. They work under the assumption that meaningful conclusions can be drawn between the number of individual performances or attributes and the customer satisfaction. Due to the ability to deliver insight into single aspects of customers’ judgments and firm performance as well as serving as a measure for overall satisfaction, multi-dimensional methods are the preferred and most widely used approaches in the field of customer satisfaction measurement. Table 2 provides a summary and evaluation of the presented methods which demonstrates the superiority of multi-dimensional measures.

illustration not visible in this excerpt

Table 2: Comparative Assessment of the Customer Satisfaction Measurement Approaches

(Based on Homburg, C., Rudolph, B. (1995), p. 47.)

Within the category of multi-dimensional methods, it is also possible to distinguish the methods by their timing of the questioning^[47]. Because of their simpler nature, one-dimensional methods almost exclusively use the ex-post variant.

2.3.5 Ex Ante/Ex Post vs. Ex Post Measurement

Measuring customer satisfaction ex ante/ex post means comparing the a priori inquired expectations to an a posteriori measurement of their fulfillment. Often in this context, a gap analysis is conducted to reveal the differences between expectations and fulfillment. This method, whose approach orientates itself strongly on the SERVQUAL approach from the measurement of service quality^[48], has received a lot of criticism in recent time and loses significance (Chapter 3.2.1). Especially due to the extremely high efforts it requires, this kind of measurement is applied only very rarely in practice. An additional problem is the fact that customers already need to have expectations about the product or service and that the expectations can become more concrete during the consumption as well^[49].

In contrast to the previous approach, an ex post measurement only applies an after the fact survey. There are two possibilities for this type of measurement. On the one hand, expectations and fulfillment can be assessed separately, similar to the ex ante/ex post approach. The main difference is the fact that both categories are surveyed in the same questionnaire. In doing so, the expectations are typically collected before the fulfillment. For this type of measurement, the disadvantages for the ex ante/ex post approach stated above remain. In this case, though, instead of requiring two surveys, only the number of questions doubles. This prevents other questions from being asked and increases the total costs of the customer satisfaction survey. The dual scale also demands high level of judgment willingness and ability from the consumers^[50].

The satisfaction measurement can, on the other hand, also be conducted on the base of a direct satisfaction judgment. At this juncture, a preliminary survey of the expectations is neglected in favor of simplicity. Nevertheless, this approach is very reliable and the most commonly used method to date because it directly captures the satisfaction instead of the extent of (dis)confirmation of expectations^[51]. Figure 4 shows possible questionnaires for the three types of measurements presented above.

illustration not visible in this excerpt

Figure 4: Examples of Ex Ante/Ex Post and Ex Post Surveys of Customer Satisfaction

(Own illustration, based on Scharnbacher, K., Kiefer, G. (1998), p. 24.)

3 Subjective Measurement Methods

After having shown that objective methods are not appropriate for measuring a complex construct like customer satisfaction, the following chapter focuses on the most popular and scientifically backed attribute-specific methods. Using the overall theoretical framework introduced Chapter 2, the primary goal of each method is presented. In a next step, their strengths and weaknesses are discussed. This chapter concludes with a comparison and evaluation of the presented methods.

3.1 Event-Specific Measures

Relying on measuring customer satisfaction by using events requires an identification of possible touchpoints first. These touchpoints function as an interface of the brand, service, or product with the customer and, thus, should always be including customers when trying to identify them. The most widely applied method in this context is blueprinting^[52]. It clearly visualizes the “line of visibility” which means separating internal activities from those the customer comes in contact with. Blueprinting is used for visualizing all processes in their temporal progression. Based on this, the actual events as well as the customers’ perception of those events have to be identified. One distinguishes between routine and non-routine events. In most cases, the focal point of the investigation is the non-routine events, since they produce enduring impressions in the customers’ memory. The Critical Incident Technique and the Frequency Relevance Analysis for Problems are the two most suitable methods for event-specific measures.

3.1.1 Critical Incident Technique

The Critical Incident Technique (CIT) is a primarily qualitative, inductive, and active method for the ascertainment of enduring customer impressions. It was developed in the 1950s by J.C. Flanagan and was adapted for the measurement of customer satisfaction in the 1980s by M.J. Bitner et al. They define critical events as “specific interactions between customers and service employees that are especially satisfying or dissatisfying”^[53]. The central idea lies in collecting and evaluating extreme encounters that often live on as stories, producing positive and negative word-of-mouth behavior. The events are surveyed with the help of personal interviews with standardized, but open questions^[54]. The test person is required only to report, not to interpret. The customer’s inherent expectations as well as perceived weaknesses in the processes and the personnel are supposed to be extracted by the interviewer without bias^[55]. Following the interviews, a multi-level analysis classifies the events as either positive or negative ones and sorts them into different problem categories by their cause. The results of the CIT can be illustrated in a table by showing the respective frequency of the corresponding categories with their positive or negative value.

[...]

---

^[1] Peterson, R., Wilson, W. (1992), p.61.

^[2] Swan, J.E., Trawick, I. F. (1993), p 30.

^[3] Oliver, R. (1997), p. 99; also numerous papers, especially Bolton, R., Drew, J. (1991); Cadotte, E. et al. (1987); Fournier, S., Mick, D. (1999); Spreng, R. et al. (1996).

^[4] Woodruff, R. et al. (1983), p. 296.

^[5] Szymanski, D., Henard, D. (2001), p. 26.

^[6] Martensen, A. et al. (2000), p. 547; also Kristensen, K. et al. (1999).

^[7] Fournier, S., Mick, D. (1999), p. 6.

^[8] Spreng, R. et al (1996), p. 15; also Hermann, A., Johnson, M.D. (1999); Stauss, B. (1999).

^[9] Cadotte, E. et al. (1987), p. 306; also LaTour, S., Peat, N. (1979).

^[10] Tse, D.K., Wilson, P.C. (1988), p. 209.

^[11] Johnson, M.D., Fornell, C. (1991), p. 10-12.

^[12] Tse, D.K., Wilson, P.C. (1988), p. 205.

^[13] Churchill, G.A., Surprenant, C. (1982), p. 492.

^[14] Parasuraman, A. et al. (1985), p. 43.

^[15] Bitner, M.J., Hubbert, A.R. (1994), p. 75.

^[16] Giering, A.(2000), p. 9.

^[17] Westbrook, R., Oliver, R.L. (1991), p. 84.

^[18] Wirtz, J., Bateson, J. (1999), p. 56; also Fornell, C. et al. (1996); Dubé, L., Morgan, M. (1998).

^[19] Richins, M. (1997), p. 132; also Mooradian, T., Olver, J. (1997).

^[20] Homburg, C. et al. (1999), p. 179.

^[21] Anderson, E.W., Sullivan, M.W. (1993); Fornell, C. et al. (1996); LaBarbera, P., Mazursky, D. (1983); Mittal, V. et al. (1998, 1999).

^[22] Oliva, T.A. et al. (1992), p. 84.

^[23] Mittal, V., Kamakura, W.A. (2001), p. 140.

^[24] Keiningham, T.L. et al (2003), p. 37; also Müller, W., Riesenbeck, H.-J. (1991).

^[25] Mittal, V., Kamakura, W.A. (2001), p. 140; also Agustin, C., Singh, J. (2005), Bowman, D., Narayandas, D. (2001).

^[26] Oliva, T.A. et al. (1992), p. 87; also Burmann, C. (1991); Hermann, A., Johnson, M.D. (1999).

^[27] Anderson, E.W. (1996), p. 272.

^[28] Anderson, E.W. et al. (1994), p. 61.

^[29] Rust, R.T. et al. (2002), p. 14.

^[30] Anderson, E.W. et al. (2004), p.181.

^[31] Gruca, T.S., Rego, L.L. (2005), pp. 120ff.

^[32] Fischer, M. et al. (2001), p. 1181.

^[33] Andreasen, A.R. (1982), pp. 183ff.

^[34] Yeung, M.C.H. et al. (2002), p. 32.

^[35] Töpfer, A., Greff, G. (2000), pp. 35ff., 66ff.

^[36] Lingenfelder, M., Schneider, W. (1991), p. 30.

^[37] Stauss, B. (2000), pp. 324ff.

^[38] Bitner, M.J., Hubbert, A.R. (1994), pp. 76ff.

^[39] Stauss, B., Hentschel, B. (1995), p. 116.

^[40] Scharnbacher, K., Kiefer, G. (1998), pp. 23ff.

^[41] Scharnbacher, K., Kiefer, G. (1998), p. 20.

^[42] Detailed in Homburg, C., Fürst, A. (2005); Stauss, B., Seidel, W. (2007).

^[43] Lingenfelder, M., Schneider, W. (1991), p. 30.

^[44] Homburg, C., Rudolph, B. (1995), p. 46.

^[45] Yi, Y. (1990), p. 72.

^[46] Standop, D., Hesse, D.-W. (1985), p. 18.

^[47] Scharnbacher, K., Kiefer, G. (1998), p. 23.

^[48] Parasuraman, A. et al. (1988): p. 12.

^[49] Hentschel, B. (2002), p. 117.

^[50] Hentschel, B. (2000), p. 311.

^[51] Scharnbacher, K., Kiefer, G. (1998), p. 26.

^[52] Stauss, B., Weinlich, B. (1996), p. 53.

^[53] Bitner, M.J. et al. (1990), p. 73.

^[54] Stauss, B., Hentschel, B. (1990), p. 241.

^[55] Bitner, M.J. et al. (1985), p 49.